7 text
MoonshotAI's coding-focused Kimi model for end-to-end software engineering.
Zhipu AI's most capable model with a major leap in coding and long-horizon task performance.
NVIDIA's efficient 30B MoE Nemotron model with 3B active parameters.
NousResearch's flagship large-scale reasoning model based on Llama 4.
OpenAI's budget-friendly reasoning model — fast and surprisingly capable.
July 2025 update to Qwen3 235B A22B with improved capabilities.
Google's open multimodal model with 128K context and 140+ language support.